# 1. Introduction

We are currently witnessing the dynamic development of small renewable micro-energy sources (RES), including mainly photovoltaic (PV) plants. There are several reasons for the increased interest in the deployment of micro-plants. The adoption of the RES Act [^{1}], which has introduced the concept of the prosumer, allowing their generation of electricity for their own needs, using the power system as virtual electricity storage, and laying down the rules for the energy balance settlement may be considered as a main reason. Another important reason is the measurable reduction of the PV plant installation cost in the last years, allowing for a real return on investment in a few years. Moreover, government mechanisms were launched to support the development of PV plants, which have a tangible financial benefit. For small PV installations in the range from 2 kW to 10 kW, the “Mój prąd” [My electricity] program [^{2}] is currently in operation, whereby a PLN 5,000 (ca. EUR 1,100) non-returnable subsidy is available. The deployment of a PV installation can also benefit from a tax credit for thermal upgrading [^{3}].

Energy-related initiatives within the European Union are focused on increasing energy efficiency, which goal is the rationalization of energy consumption. This is reflected in the Polish legislation in the form of the Act of May 26, 2016, on energy efficiency [^{4}].

The implementations proposed in the Act to improve energy efficiency relate to reduction of losses in the distribution of electricity, in transmission facilities and in grids managed by distribution system operators (DSOs). The use of RES micro-sources to reduce losses incurred in HV/MV substations may be in line with the goals of increasing the energy efficiency of DSOs. This paper describes how PV installations can be used to fulfil the electricity demand of HV/MV substations. A real HV/MV station was used in these considerations.

# 2. HV/MV substation load and losses

An HV/MV substation is equipped with auxiliary systems which consumes electrical energy and they are fed by auxiliary transformers (MV/LV). Their energy demand can be interpreted as a part of energy losses in the substation which include energy losses in power transformers (HV/MV). Therefore, the following considerations take into account the substation auxiliaries energy demand, and energy losses in the auxiliary transformers and energy losses in the power transformers.

The analyses were based on the actual parameters and measurements of the selected 110 kV/15 kV transformer substation. There are two auxiliary transformers, 160 kVA each, in the considered substation. The transformers are older units characterized by load losses of 1.9 kW and relatively large no-load losses of 3 kW. They are used to supply the substation’s auxiliaries. The active and reactive power load profiles are shown in Fig. 1.

Fig. 1. Annual active and reactive power load profile of the auxiliary transformers

The load profiles of auxiliary transformers were used to determine the annual profiles of energy losses in these transformers, which are shown in Fig. 2 and listed in Tab. 1. It should be noted that the small load of both transformers in the summer period results in total power losses close to the no-load losses. In turn, in the case of the PW 2 transformer, such a situation occurs practically all year round. Therefore, the decision to replace transformers with new generation units with better parameters, including lower rated power and reduced no-load losses, may be economically justified.

Fig. 2. Annual profile of energy losses in auxiliary transformers

Tab. 1. Energy consumption for auxiliaries supply and energy losses in auxiliary transformers in the HV/MV substation

It can be noticed that the share of transformer energy losses in the auxiliaries load is relatively large and, depending on the month, ranges from 19.3% to 41.7%, with an average annual share of 29.3% (Tab. 1). This is mainly due to the previously mentioned large no-load losses of auxiliary transformers.

The aggregate annual electricity demand, being the sum of the energy consumed by the station's auxiliaries and the energy losses in the auxiliary transformers, amounted to 201.3 MWh, which translates into a cost of PLN 62,666/year, assuming the electricity cost of PLN 311/MWh.

The 110 kV/15 kV transformers in the analysed substation are 25 MVA units with rated no-load losses of 10.7 kW and rated load losses of 134 kW. The transformer load changes throughout the year (Fig. 3) and is slightly higher in winter. The average annual load of the TR1 transformer is ca. 28%, and the TR2 transformer 17%.

Fig. 3. Annual load profile of the 110 kV/15 kV transformers

From the load profiles, the power losses profiles can be determined for both transformers, and then the energy losses of the transmission between the 110 kV and 15 kV grids can be calculated (Fig. 4 and Tab. 2). The energy losses are the basis for estimating the cost of energy transformation in the power transformers, which the operator incurs in connection with supplying the MV grid from the HV/MV substation. The aggregate annual energy losses in both transformers are 344.8 MWh, which translates into the cost of losses of PLN 107,232.8/year.

Fig. 4. Annual profiles of energy losses in the 110 kV/15 kV transformers

Tab. 2. Aggregate energy losses in the 110 kV/15 kV transformers

# 3. HV/MV substation potential in terms of RES installations

For the HV/MV substation, possible areas where a PV installation can be located were selected. It should be noted that the selected areas were not consulted with the station owner for the PV installation deployment feasibility. The selection of individual locations was made based on satellite images available from Google Maps. Six areas were identified on the substation site, and their approximate dimensions are shown in Fig. 5.

Fig. 5. The HV/MV substation site

Three aggregate PV plant power options were considered:

- Option W1 – Area 1 and Area 2 In this option, the PV panels should not be covered by the shadows of other objects in the substation. The winter months may be an exception.
- Option W2 – Area 1, Area 2, and Area 6 This option includes the areas from the W1 option, and also Area 6, which should not be covered by the shadows of other objects in the station either.
- Option W3 – all areas. In this option, areas 3, 4, and 5 will be periodically obscured by the shadows cast by other objects in the substation, e.g. HV poles or wires, which will reduce the average annual electricity output.

For simplicity, in further considerations, the power generated by the PV panels is assumed as if they were not obscured by the shadows of other objects, which is correct in the W1 and W2 options. In the W3 option, it should be remembered that the actual output power will be lower than that presented.

An RSM60-6-295M-315M/5BB panel [^{5}] was adopted for the analyses, a 310 W unit with dimensions of 0.992 m × 1.65 m, which were taken into account when calculating the number of panels that could be installed in each area. The panels' manufacturer declares that over 25 years of operation their efficiency should be at least 80% of the initial value.

Five of the selected areas are ground locations, and one area is the roof of a substation building, which is inclined at an angle of approx. 6°. The roof dimensions are ca. 12 m × 29 m. The building's location is advantageous, as its long side faces south with a slight west orientation.

The PV panels' deployment on the roof's one, the south-facing slope was assumed. It was assumed that the panels on the roof will be installed with the longer side horizontally. The reason was the roof's low slope, and thus the likelihood of snow retention after rainfall. It is intended to use PV panels with bypass diodes, which for economic reasons are usually mounted on the shorter side of the panel, which allows the use of fewer diodes. In such a situation, if the upper part of the panel is exposed from the snow, a current begins to flow in this part causing it to heat up. This in turn should melt the snow and further expose the panel.

Other potential areas of panel installation require an on-ground mount. It was assumed that the panels would be mounted on special supporting structures, keeping a certain distance from the bottom edges of the panels to the ground (ca. 0.5 m). This allows avoiding mud splatter of the lower parts of the panels when raining. The optimal slope of the panels relative to the ground for the latitude of Poland is the range from 25° to 40°. The angle of inclination of the panels determines the spacing of individual rows of panels. It was assumed that the individual rows of panels should not shade each other at the zenith throughout the year. Thus, the greater the panel inclination angle, the greater the distance between the rows will be. To obtain the smallest distance between the rows, the inclination angle was assumed to be 25°. At the same time, it was assumed that the three rows of panels with horizontally-aligned longer sides would be installed on supporting structures (Fig. 6).

Fig. 6. Distance between rows of PV panels

Taking the above into account, the assumed distance between the rows of panels is [^{6}]:

where: d – sum of the three lengths of the shorter side of the panel, β – module inclination angle (25°), α = 90° – latitude – 23°27’.

Generation profiles were developed for each of the above variants. For this purpose, the “Photovoltaic Geographical Information System” website [^{7}] was used. The resulting output power profiles take into account the respective inclination of the panels and their deviation from the south direction. Using these profiles, the annual electricity output of the PV plant in each option was estimated (Fig. 7).

Fig. 7. The annual energy output of the PV panels in various options

The PV plant will be connected to the auxiliary transformer's low voltage side. It was assumed that it would be the PW 1 transformer, which is more loaded. Thus, more energy for the auxiliaries will be delivered by the PV plant directly on the low voltage side. The power in the PW 1 transformer flows in both directions, after connecting the PV plant, mainly from spring to autumn (Fig. 8, 9, 10). In this period, the power flow direction of changes throughout the day. In the daytime, it flows towards the MV grid, and at night towards the LV grid (auxiliaries).

Fig. 8. Annual load profile of the transformer TR PW 1 with PV panels. Option W1 (negative values indicate the power input to the grid)

Fig. 9. Annual load profile of the transformer TR PW 1 with PV panels. Option W2 (negative values indicate the power input to the grid)

Fig. 10. Annual load profile of the transformer TR PW 1 with PV panels. Option W3 (negative values indicate the power input to the grid)

It should be noted that in the W3 option, the installed PV capacity is greater than the rated power of the auxiliary transformer (PPV W3 = 192.82 kW) > (PTR PW1 = 160 kW), nevertheless, the PV plant's effective power output will be lower and below its rated capacity (Fig. 10).

Electricity generation in a photovoltaic plant increases the power flow through the auxiliary transformer. The power flow is much higher than that resulting from the substation's auxiliary consumption, however, due to the transformer's relatively high no-load losses, the aggregate energy losses in each month of the year do not differ significantly after connecting the PV plant, regardless of the option considered (Fig. 11). Winter months are characteristic here, for which these losses are very close to each other.

Fig. 11. Annual energy losses in the auxiliary transformers in various options

When a PV plant is connected, the energy balance of the substation's supply system may be positive or negative depending on the month and the considered option (Fig. 12). A positive electricity balance (energy consumption) in each month is obtained for option W1, which is characterised by the lowest installed PV plant capacity (68.2 kW). In the other two options, in the months from May to September, a negative balance is obtained, which means that in total, monthly, energy is fed into the grid.

Fig. 12. Energy balance of the auxiliary power supply system, including energy losses in the auxiliary transformers, in various options (negative values mean energy fed into the grid)

The annual demand of the HV/MV substation for electricity needed to supply its auxiliaries, without losses in the power transformers (TR1 and TR2, amounts in the consideration period to 201.3 MWh. The PV plant's addition reduces this demand. The best effect, close to balancing the demand on an annual basis, is obtained in the W3 option.

The demand of the substation for electricity, besides its auxiliaries' supply, should also include energy losses in the 110 kV/15 kV transformers (344.8 MWh). In this case, the aggregate annual electricity demand of the substation in the period under consideration is 546 MWh. Therefore, taking into account the maximum PV plant power in the W3 option (192.82 kW), there remains over 66.7% of electricity consumed at the substation in the annual balance (Fig. 13). This proves that the addition of a ca. 200 kW PV plant does not cover even 50% of the substation's electricity demand.

Fig. 13. Energy balance for the entire HV/MV substation in various options

Moreover, Fig. 14 includes the information on how much the CO2 emission will reduce in each option, as a result of electricity generation by the PV power plant in the first year of its operation. In the following years, this reduction will gradually decrease along with decreasing the PV panels' efficiency. These values have been determined with the consideration of the CO2 emission factor for end-users of electricity of 765 kg/MWh [^{9}].

Fig. 14. CO_{2} emission reduction as a result of electricity generation in a PV plant

# 4. Cost-effectiveness analysis of the RES source deployment in the HV/MV substation

The analyses presented above have shown that the proposed PV plant sites do not ensure the annual balancing of the HV/ MV substation's electricity demand. Option W3, which includes all these sites, is close to it. It should be remembered that the considerations do not include the temporary obscuring of some panels by the shadows of other objects at the substation, which will reduce the annual electricity output of the PV plant in this option.

The three PV plant capacity options proposed above were analysed for cost-effectiveness. The following assumptions were adopted in the considerations:

- electricity price increases every year by 1% from PLN 311/ MWh to PLN 394.89/MWh in the 25th year of the project
- total PV plant deployment cost is PLN 5,000/kW1, hence the PV plant deployment costs in various options are following: W1 – PLN 341,000, W2 – PLN 675,800, W3 – PLN 964,100
- annual servicing costs are 50 PLN/kW2, therefore annual servicing costs in individual variants, starting from the second year of the investment, are as follows: W1 – 3,410 PLN, W2 – 6,758 PLN, W3 – 9,641 PLN
- SolarEdge SE27.6K inverters with a rated power of 27.6 kW [8] were used
- the W1 variant uses 2 inverters, the W2 variant uses 4 inverters, and the W3 variant uses 6 inverters, which results from the PV plant output power profiles
- after 15 years of operation, all inverters shall be replaced at PLN 8,800 for each inverter.

The energy produced in PV plants balances the electricity used at a given moment in the HV/MV substation only during a specific period of the day – when the solar conditions are suitable for the plant's operation. The approach presented in the paper is based on balancing the annual electricity demand of the HV/MV substation and the annual electricity output of the PV plant.

Under the above assumptions, with a simple investment return period, in each of the considered options, the capital expenditure will be recovered after 21 years of the PV plant operation (Fig. 15).

Fig. 15. The cumulative cost of electricity produced in the plant

The simple payback period application does not answer as to whether investing money in another project will be more profitable. To check the effectiveness of the investment, the commonly used NPV (Net Present Value) ratio was used:

where: CFi – net cash flow expected in the year i, N0 – initial capital expemditure, k – discount rate.

The investment effectiveness was checked for various discount rates: 1.5%, 2%, 3%, 5%, and 8%, with consideration of the replacement of inverters in the 15th year of the project. The results of the analysis are shown in Fig. 16. A negative value of the NPV ratio means that the investment is unprofitable. A positive value of this ratio is obtained for a discount rate of not more than 1.5%. Thus, the profitability of the considered investment will depend on the discount rate used by the operator for this type of project.

Fig. 16. NPV ratio for the considered PV plant options with various discount rates

The return on investments can be more precisely verified using the Net Present Value Ratio (NPVR), defined as the reference of the NPV ratio to investment outlays. With this, it can be checked in which option the investment effectiveness is best (Fig. 17). Over the 25 years considered, the best profitability is obtained for options W2 and W3, at discount rates of 1% and 1,5%.

Fig. 17. NPVR ratio for the considered PV plant options with various discount rates

The above considerations also show that despite the continuous decrease in the cost of this type of capex projects, it is still difficult to achieve the investment effectiveness without external support. Nevertheless, reducing the PV installation price per 1 kW leads to a reduction in the time necessary to recover the investment outlay (Tab. 3).

Tab. 3. The simple return period of the 68 kW PV plant (W1) [years]

Summing up, it should be remembered that the profitability of an investment in a PV power plant installed in an HV/MV substation strongly depends on several of the following variables:

- average price per 1 kW of installed power plant capacity
- increase in the price of electricity over the assumed project duration
- discount rate
- annual maintenance costs
- investment replacement costs.

# 5. Summary

HV/MV substations are the power system components, where electricity is consumed in connection with their operation. The electricity demand was considered here as the sum of electricity consumed by the loads (auxiliaries) installed in the substation, energy losses in auxiliary transformers, and energy losses in power (HV/MV) transformers. The total annual electricity demand of the considered substation is 546 MWh, which, assuming the electricity price of PLN 311/MWh, gives the annual energy purchase cost PLN 170,000. The annual demand consists of the demand for auxiliaries (201.3 MWh) and the coverage of energy losses incurred in power transformers (344.8 MWh). It should be noted that the energy losses incurred in the power transformers in the substation are definitely greater than the total energy required to supply the station's auxiliaries. Hence the conclusion that, if possible, they should be included in the substation's electricity demand. Moreover, attention should be paid to the high no-load losses of the auxiliary transformers in the substation. Currently, manufactured transformers of this type have a no-load loss of 1 kW. Replacing one transformer would save 35 MWh annually, which translates into savings of PLN 10,885/year. With the price of a new transformer of the same power with the lowest no-load losses (130 W) and load losses (1,620 W) of PLN 25,000, replacement of the transformer would be repaid within 2.5 years.

For a PV plant, the location is important, which should be free from the shadows cast by other objects in the substation. Considering the towers of the HV lines incoming to the substation and situated in it, not all potential PV plant locations will be free from temporary shading on some PV panels, which will reduce the plant's annual average electricity output.

The economic analysis has shown that, in the case of the HV/ MV substation, the location considerations prevent covering its annual average energy needs. Nevertheless, regardless of the considered PV plant capacity option, the profitability always depends on the adopted discount rate. The plant capacity is not decisive here, and therefore the profitability is in a sense independent of the HV/MV substation (if a site that is not temporarily shaded is available). Considering the 25-year lifetime of a PV plant, at an electricity price of PLN 5,000/kW, the investment is profitable for a discount rate of 1.5% or lower. Considering the simple payback period for each of the three options considered, the investment pays off after 21 years. The time of return on investment is significantly shorter at lower prices per 1 kW, e.g. at the price of 3,500 PLN/kW and its 3% annual increase the simple payback period is reduced to 12 years.

When considering the business aspect of profitability of the deployment of a PV plant in an HV/MV substation, it should be noted that these considerations are universal and independent of the specific substation. Therefore, they can be applied to any HV/MV substation. The considerations show that the annual electricity demand of an HV/MV substation is relatively large in relation to the possibility of satisfying it from a RES source, where the main obstacle will be the availability of the source's appropriate location at the substation.

The idea of using RES sources to cover the electricity needs of HV/MV substations is, by all means, correct and is very clearly in line with the European Union's pro-climate initiatives aiming at climate-neutral Europe by 2050. Depending on the adopted option (from W1 to W3), with a PV installation in one HV/MV substation, the annual CO2 emission can be reduced from 50 tons to 146 tons. The RES source deployment in HV/MV substations serves to increase the energy efficiency of DSOs by striving for an annual average balance of electricity necessary for the HV/ MV station operations. Therefore, it seems reasonable to lobby for the introduction of mechanisms to support the use of RES sources to increase the energy efficiency of HV/MV substations.